Why is my site in Google but not in AI answers?

Difference one: rendering

The largest and most common cause. Googlebot executes JavaScript on a second pass — slowly and selectively, but it does. Most AI crawlers do not render at all.

A site built with client-side rendering therefore presents two entirely different faces: a complete page to Google, and an empty container to almost every AI crawler. It can rank respectably while being invisible to every assistant, and nothing in your search analytics would ever hint at it.

curl -sL https://yoursite.com | grep -i "a distinctive sentence from your page"

If that returns nothing, this is your answer, and it is one change rather than a strategy.

Difference two: pages versus passages

Google ranks documents. A person clicks and finds the relevant part themselves.

An assistant extracts a passage and presents it as the answer. That requires a self-contained span of text that makes sense removed from its surroundings. A page that builds an argument over eight paragraphs is excellent for a reader and offers nothing liftable.

Difference three: different crawlers, different permissions

Allowing Googlebot says nothing about the others. Each operator uses its own tokens, and a robots.txt permitting Google can block every AI crawler — often unintentionally, via a blanket rule or a copied snippet.

The specific trap: a catch-all Disallow: / with a named Allow group for Googlebot. Google is fine. Every crawler without its own group is blocked, which is most of them.

Difference four: no accumulated history

Your Google position reflects years of signals — links, engagement, domain history. An answer engine choosing a source for one question weighs the page in front of it far more heavily. That cuts both ways: established sites lose some of their structural advantage, and a well-structured page from a smaller site can be chosen over a famous one.

Working through it

1. curl -sL your page and confirm your content is in the delivered HTML. This is the single highest-yield check.
2. Read your robots.txt and confirm the AI search crawlers have permission, by name.
3. Check your logs for their requests — and specifically the status codes they receive, since a WAF can block what robots.txt permits.
4. Restructure your key pages so each answers a specific question directly beneath a heading that asks it.
5. Add clear attribution: a named author, an organisation, a date.

The first three are mechanical and take under half an hour between them. They account for most cases where a site ranks in Google and appears nowhere else.

When the mechanics are already right

If you are crawled, indexed and readable and still not cited, you are competing on merit — and the competitive question is narrower than in search. Look at who is cited for your topic and what their pages do: usually a direct answer near the top, original detail rather than a summary of others, and visible authorship. That is a content project rather than a technical one, and it is a much better position to be in.